CN108417221A - The digital walkie-talkie acoustic code type detection method of fused filtering is recombinated based on signal two dimension - Google Patents
The digital walkie-talkie acoustic code type detection method of fused filtering is recombinated based on signal two dimension Download PDFInfo
- Publication number
- CN108417221A CN108417221A CN201810074257.XA CN201810074257A CN108417221A CN 108417221 A CN108417221 A CN 108417221A CN 201810074257 A CN201810074257 A CN 201810074257A CN 108417221 A CN108417221 A CN 108417221A
- Authority
- CN
- China
- Prior art keywords
- signal
- acoustic code
- code signal
- fused filtering
- dimension
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000001914 filtration Methods 0.000 title claims abstract description 40
- 238000001514 detection method Methods 0.000 title claims abstract description 34
- 238000000034 method Methods 0.000 claims abstract description 39
- 230000006798 recombination Effects 0.000 claims abstract description 20
- 238000005215 recombination Methods 0.000 claims abstract description 20
- 238000012360 testing method Methods 0.000 claims abstract description 16
- 238000012706 support-vector machine Methods 0.000 claims description 4
- 230000004927 fusion Effects 0.000 abstract description 7
- 238000000605 extraction Methods 0.000 description 5
- 230000000694 effects Effects 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000002790 cross-validation Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000007500 overflow downdraw method Methods 0.000 description 1
- 230000000644 propagated effect Effects 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
- G10L19/04—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis using predictive techniques
- G10L19/16—Vocoder architecture
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/08—Speech classification or search
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
- G10L19/04—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis using predictive techniques
- G10L19/26—Pre-filtering or post-filtering
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
- G10L25/03—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters
Landscapes
- Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Complex Calculations (AREA)
- Telephonic Communication Services (AREA)
Abstract
The invention discloses a kind of digital walkie-talkie acoustic code type detection methods recombinating fused filtering based on signal two dimension, include the following steps：2D signal recombination is carried out to one-dimensional acoustic code signal, fused filtering processing is carried out to the two-dimentional acoustic code signal after recombination, model training is carried out according to the feature that attribute reduction obtains to the two-dimentional acoustic code signal Jing Guo fused filtering, obtains svm classifier model；2D signal recombination and fused filtering processing are carried out to digital walkie-talkie acoustic code sample of signal to be detected, use svm classifier model, to fused filtering, treated that 2D signal carries out test detection according to the feature that attribute reduction obtains, decision level fusion is carried out to the test testing result of each 2D signal using ballot method, obtains final acoustic code signal type detection result.One-dimensional acoustic code signal two dimension is improved the precision of acoustic code signal type detection by the present invention using the method for signal two dimension recombination so as to preferably extract acoustic code signal characteristic.
Description
Technical field
The invention belongs to digital processing fields, and in particular to a kind of number recombinating fused filtering based on signal two dimension
Intercom acoustic code type detection method.
Background technology
Digital walkie-talkie is the intercom being designed using digital technology.Digital walkie-talkie by digitization of speech signals,
It is propagated in the form of digital coding, whole modulation in intercom transmission frequency are number.Compared with analog walkie-talkie, number is right
Say machine have strong antijamming capability, speech quality is good, frequency efficiency is high, security performance is good, support data service, be convenient for
The features such as zero defect relays, is widely used.The sound of digital signal in some specific places generally require to digital walkie-talkie
Code type is detected, further to identify the content of digital walkie-talkie transmission.
The key of acoustic code signal type detection is the feature extraction of signal.Existing acoustic code signal type detection algorithm is equal
It is that acoustic code type detection is carried out to one-dimensional acoustic code signal, since one-dimensional acoustic code signal is easy affected by noise, and class internal difference
Different apparent, the feature of input signal is unable to get effective extraction, and the precision and stability of acoustic code signal type detection algorithm has
It waits improving.
Invention content
The purpose of the present invention is to provide a kind of digital walkie-talkie acoustic code types recombinating fused filtering based on signal two dimension
Detection method improves the precision and stability of digital walkie-talkie acoustic code type detection, is suitable for the high-precision of small sample acoustic code signal
Acoustic code type efficient detection.
Realize that the technical solution of the object of the invention is：A kind of digital walkie-talkie sound recombinating fused filtering based on signal two dimension
Code type detection method, includes the following steps：
Step 1,2D signal recombination is carried out to one-dimensional acoustic code signal, obtains two-dimentional acoustic code signal；
Step 2, the two-dimentional acoustic code signal after recombination is pre-processed using fused filtering method；
Step 3, using support vector machines as grader, to the two-dimentional acoustic code signal Jing Guo fused filtering according to attribute about
The feature that letter obtains carries out model training, obtains svm classifier model；
Step 4,2D signal recombination is carried out to digital walkie-talkie acoustic code sample of signal to be detected and fused filtering is handled；
Step 5, obtained svm classifier model is trained using step 3, to fused filtering treated two-dimentional acoustic code signal root
The feature obtained according to attribute reduction carries out test detection, using ballot method to the test testing result of each two-dimentional acoustic code signal into
Row decision level fusion obtains final acoustic code signal type detection result.
Compared with prior art, the present invention its remarkable advantage is：(1) present invention will using the method for signal two dimension recombination
The significantly one-dimensional acoustic code signal two dimension of all kinds of interior differences improves acoustic code signal so as to preferably extract acoustic code signal characteristic
The precision of type detection；(2) influence of noise, and the method for using attribute reduction are reduced using the method for fused filtering, from
Validity feature is selected in each group feature of acoustic code signal so that acoustic code signal type detection is more reliable and more stable；(3) pass through ballot
Method carries out decision level fusion, further improves accuracy of detection.
Description of the drawings
Fig. 1 is the digital walkie-talkie acoustic code type detection method flow chart that fused filtering is recombinated based on signal two dimension.
Specific implementation mode
The present invention solves the problems, such as that difference is excessive in one-dimensional signal class by the two dimension recombination of signal, using fused filtering
Method pre-processes acoustic code signal, influence of the removal noise to acoustic code signal.The method of application attribute yojan is from advance
A variety of validity features are selected to be analyzed in digital walkie-talkie two dimension acoustic code signal after reason, then by the acoustic code signal two of extraction
Dimension acoustic code signal characteristic is input to training in SVM classifier and obtains svm classifier model, and using this disaggregated model to passing through two
The test signal of dimension recombination and fused filtering carries out category analysis.A kind of decision level fusion method is finally used, to prediction result
It is handled, obtains the corresponding label of each test signal.The inventive method can be obtained in the case where training sample is smaller
In high precision, stablize, the acoustic code signal detection effect of robust.
In conjunction with Fig. 1, a kind of digital walkie-talkie acoustic code type detection method recombinating fused filtering based on signal two dimension, including with
Lower step：
Step 1, to one-dimensional acoustic code signal carry out 2D signal recombination, by the one-dimensional acoustic code signal comprising D byte according to
T rows are divided into, two-dimentional acoustic code signal is obtained；
Step 2, the two-dimentional acoustic code signal after recombination is pre-processed using fused filtering method；
Step 3, using support vector machines as grader, to the two-dimentional acoustic code signal Jing Guo fused filtering according to attribute about
The feature that letter obtains carries out model training, obtains svm classifier model；
Step 4,2D signal recombination is carried out to digital walkie-talkie acoustic code sample of signal to be detected and fused filtering is handled；
Step 5, obtained svm classifier model is trained using step 3, to fused filtering treated two-dimentional acoustic code signal root
The feature obtained according to attribute reduction carries out test detection, using ballot method to the test testing result of each two-dimentional acoustic code signal into
Row decision level fusion obtains final acoustic code signal type detection result.
Further, step 1 is specially：
Assuming that there is N number of different sample：X={ x1,x2,…,xN|xi∈RD, i ∈ [1,2 ..., N] }, wherein D is one-dimensional sound
The data volume of each sample of code signal, xiFor sample, N is sample number；Set Y={ the y of one-dimensional acoustic code signal label1,y2,…
yi…,yN| i=1,2 ..., N }, yiFor the label of i-th of one-dimensional acoustic code signal；
One-dimensional sample information is converted into two dimensional sample information, transformed sample can be expressed as：X'={ x11,
x12,…,xNT|xij∈RD/T,i∈[1,2,…,N],j∈[1,2,…,T]}
Wherein xijFor the jth row of i-th of two-dimentional acoustic code signal, T is the row for the 2D signal that each one-dimensional signal is divided into
Number after the completion of division, obtains N*T rows, the two dimensional sample data of D/T row.
Preferably, in step 1 by the one-dimensional acoustic code signal comprising 189 bytes according to being divided into 7 rows.
Further, fused filtering is recombinated to two dimension by the method for acoustic code signal strength mean filter in step 2
Acoustic code signal afterwards is pre-processed；
The two-dimentional sleiding form for constructing 3*3 first, by a 3*3 neighborhood of the value of some point in two-dimentional acoustic code signal
In the statistical value of each point value replace, the point of vacancy takes zero processing, acoustic code signal strength mean filter formula as follows：
(x, y) indicates that the position of any in two-dimentional acoustic code signal, g (x, y) are the statistical value of the point, and f (x-k, y-l) is point
Value at (x-k, y-l), n are template size, k, l values be -1,0,1, W be 3*3 two-dimentional sleiding form.
The following describes the present invention in detail with reference to examples.
Embodiment
In conjunction with Fig. 1, a kind of digital walkie-talkie acoustic code type detection method recombinating fused filtering based on signal two dimension, specifically
Process is：
Step 1, this method carries out signal two dimension recombination to one-dimensional acoustic code signal first, includes 189 bytes by script
One-dimensional acoustic code signal is divided according to 27 bytes of every row.
Assuming that there is N number of different sample：X={ x1,x2,…,xN|xi∈RD, i ∈ [1,2 ..., N] }, Y={ y1,y2,…
yi…,yN| i=1,2 ..., N }, wherein D is the data volume of one-dimensional each sample of acoustic code signal, and size is 189 bytes, and N is sample
Number；Y is the set of one-dimensional acoustic code signal label, yiFor the label of i-th of one-dimensional acoustic code signal.
Since one-dimensional acoustic code signal is easy affected by noise, and difference is apparent in class, and the feature of input signal can not
Effectively extracted.In order to more fully utilize sample information, we select using the method for recombinating one-dimensional signal two dimension
To carry out signal recombination to sample.By by script comprising 189 bytes one-dimensional acoustic code signal according to 27 bytes of every row into
Row segmentation obtains the 2D signal that line number is 7 (189/27), one-dimensional sample information is converted to two dimensional sample information, after conversion
Sample can be expressed as：
X'={ x11,x12,…,xNT|xij∈RD/T, i ∈ [1,2 ..., N], j ∈ [1,2 ..., T] }, wherein T is each one
The line number for the 2D signal that dimensional signal is divided into.After the completion of division, we will obtain N*T rows, the two dimensional sample data of D/T row,
xijFor the jth row of i-th of two-dimentional acoustic code signal.
Step 2, the 2D signal after recombination is pre-processed using fused filtering method.
Acoustic code after mainly being recombinated to two dimension by the method for acoustic code signal strength mean filter during fused filtering
Signal is pre-processed.The two-dimentional sleiding form for constructing 3*3 first, one of the value of some point in two-dimentional acoustic code signal
The statistical value of each point value replaces in 3*3 neighborhoods, and this method can effectively inhibit noise.Acoustic code signal is strong during fused filtering
It is as follows to spend mean filter formula
G (x, y)=∑ (f (x-k, y-l))/n, (k, l ∈ W)
(x, y) indicates that the position of any in two-dimentional acoustic code signal, g (x, y) are the statistical value of the point, and f (x-k, y-l) is point
Value at (x-k, y-l), n are template size, k, l values be -1,0,1, W be 3*3 two-dimentional sleiding form.
Step 3, for by the pretreated two-dimentional acoustic code signal characteristic of fused filtering method, the side of attribute reduction is used
Method selects the feature that can efficiently differentiate out different classes of signal, and high-dimensional feature space is compressed to low-dimensional feature space.It will
The characteristic obtained by attribute reduction method is used as input, and using support vector machines as grader, processing is extracted
Feature, obtain svm classifier model.
In order to reduce the complexity of calculating, the accuracy rate of Classification and Identification is improved, present invention selection uses the side of attribute reduction
Method.According to the characteristic of digital walkie-talkie signal, multi-signal characteristic type is selected to be analyzed, including signal bandwidth, mean value, side
Difference, peak point, frequency point degree, normalization instantaneous amplitude absolute value mean square deviation, normalization instantaneous amplitude kurtosis, spectrum signal are inclined
Difference.All features of extraction are denoted as feature set, the method for application attribute yojan selects most effective feature from feature set,
High-dimensional feature space is compressed to low-dimensional feature space, realizes the high-precision detection of different acoustic code signal types.
Present invention application SVM classifier classifies to the acoustic code signal characteristic of extraction.SVM classifier Kernel Function selects
It selects, penalty factor and kernel functional parameter g directly affect SVM and conclude performance.The effect of kernel function is mainly by non-linear sample
Data are mapped to higher dimensional space, to realize linear classification in higher dimensional space.In order to obtain optimal classification as a result, avoiding different IPs
The influence of function pair classifying quality compares proof through overtesting and achieves higher nicety of grading using RBF kernel functions, therefore
The kernel function used in method of the present invention is RBF radial basis function.The effect of penalty factor is the attention journey characterized to outlier
Degree, kernel functional parameter g indicate kernel function radius, the two parameters we mainly use network cross-validation method (abbreviation Grid
Method) training rule of thumb selects some possible C, g values to obtain highest classification essence using k crossing method optimum selectings
The parameter of degree.Parameter C values are 100 by the present embodiment, and parameter g values are 1e-5.
Step 4,2D signal recombination is carried out to digital walkie-talkie acoustic code sample of signal to be detected and fused filtering is handled.
Step 5, for the pretreated acoustic code sample of signal of step 4 is passed through, being gone out using the method choice of attribute reduction can have
The feature of different classes of signal is distinguished to effect, and uses each feature of svm classifier model prediction acoustic code signal obtained by step 3
Corresponding classification carries out Decision fusion using ballot method to the classification results predicted by each feature of same sample, improves
Nicety of grading, and obtain the final result corresponding to sample.
This method using svm classifier model prediction acoustic code signal attribute to be detected it is brief after each feature corresponding to class
Not, Decision fusion is then carried out by method of voting, which is divided into and obtains one kind of identical decision by most features.
Enable Xt={ x1,x2,…,xT|xj∈RD, j ∈ [1,2 ..., T] } and indicate the two dimension of single acoustic code signal testing sample
Feature, Yt={ y1,y2,…,yT ',Indicate classification corresponding to each feature after single acoustic code signal testing sample attribute is brief
As a result, T ' is the dimension after attribute reduction；The expression formula that Decision fusion is carried out using ballot method is as follows：
Wherein, count (p) indicates classification p in sample label YtThe number of middle appearance, C are the possibility classification of acoustic code signal
Set.
Claims (5)
1. a kind of digital walkie-talkie acoustic code type detection method recombinating fused filtering based on signal two dimension, which is characterized in that including
Following steps：
Step 1,2D signal recombination is carried out to one-dimensional acoustic code signal, obtains two-dimentional acoustic code signal；
Step 2, the two-dimentional acoustic code signal after recombination is pre-processed using fused filtering method；
Step 3, using support vector machines as grader, the two-dimentional acoustic code signal Jing Guo fused filtering is obtained according to attribute reduction
The feature obtained carries out model training, obtains svm classifier model；
Step 4,2D signal recombination is carried out to digital walkie-talkie acoustic code sample of signal to be detected and fused filtering is handled；
Step 5, obtained svm classifier model is trained using step 3, to fused filtering treated two-dimentional acoustic code signal according to category
Property the feature that obtains of yojan carry out test detection, determined to the test testing result of each two-dimentional acoustic code signal using ballot method
Plan grade merges, and obtains final acoustic code signal type detection result.
2. the digital walkie-talkie acoustic code type detection method according to claim 1 that fused filtering is recombinated based on signal two dimension,
It is characterized in that, step 1 is specially：
Assuming that there is N number of different sample：X={ x1,x2,…,xN|xi∈RD, i ∈ [1,2 ..., N] }, wherein D believes for one-dimensional acoustic code
The data volume of number each sample, xiFor sample, N is sample number；
One-dimensional sample information is converted into two dimensional sample information, transformed sample is expressed as：X'={ x11,x12,…,xNT|xij
∈RD/T,i∈[1,2,…,N],j∈[1,2,…,T]}；
Wherein xijFor the jth row of i-th of two-dimentional acoustic code signal, T is the line number for the 2D signal that each one-dimensional signal is divided into, and is drawn
After the completion of point, N*T rows, the two dimensional sample data of D/T row are obtained.
3. the digital walkie-talkie acoustic code type detection method according to claim 2 that fused filtering is recombinated based on signal two dimension,
It is characterized in that, by the one-dimensional acoustic code signal comprising D byte according to T rows are divided into step 1, wherein D is the integral multiple of T.
4. the digital walkie-talkie acoustic code type detection method according to claim 3 that fused filtering is recombinated based on signal two dimension,
It is characterized in that, the value that the value of D is 189, T is 7.
5. the digital walkie-talkie acoustic code type detection according to claim 1 or 2 for recombinating fused filtering based on signal two dimension
Method, which is characterized in that fused filtering is after being recombinated to two dimension by the method for acoustic code signal strength mean filter in step 2
Acoustic code signal is pre-processed；
The two-dimentional sleiding form for constructing 3*3 first, will be each in a 3*3 neighborhood of the value of some point in two-dimentional acoustic code signal
The statistical value of point value replaces, and the point of vacancy takes zero processing, acoustic code signal strength mean filter formula as follows：
G (x, y)=∑ (f (x-k, y-l))/n, (k, l ∈ W)
(x, y) indicates that the position of any in two-dimentional acoustic code signal, g (x, y) are the statistical value of the point, and f (x-k, y-l) is point (x-
K, y-l) at value, n is template size, k, l values be -1,0,1, W be 3*3 two-dimentional sleiding form.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810074257.XA CN108417221B (en) | 2018-01-25 | 2018-01-25 | Digital interphone sound code type detection method based on signal two-dimensional recombination fusion filtering |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810074257.XA CN108417221B (en) | 2018-01-25 | 2018-01-25 | Digital interphone sound code type detection method based on signal two-dimensional recombination fusion filtering |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108417221A true CN108417221A (en) | 2018-08-17 |
| CN108417221B CN108417221B (en) | 2021-09-21 |
Family
ID=63126159
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810074257.XA Active CN108417221B (en) | 2018-01-25 | 2018-01-25 | Digital interphone sound code type detection method based on signal two-dimensional recombination fusion filtering |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108417221B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1064562B1 (en) * | 1998-03-17 | 2009-01-14 | QUALCOMM Incorporated | System and method for determining the position of a wireless cdma transceiver |
| US8213953B1 (en) * | 2004-05-11 | 2012-07-03 | Sprint Spectrum L.P. | Method and system for vocoder bypass based on loads in coverage areas of a wireless wide area network |
| CN102763160A (en) * | 2010-02-18 | 2012-10-31 | 高通股份有限公司 | Microphone array subset selection for robust noise reduction |
| US20160249251A1 (en) * | 2015-02-19 | 2016-08-25 | Icom Incorporated | Radio communication system, radio apparatus, radio communication method, packet generation method, and method for reproducing data from packet |
| US20160329922A1 (en) * | 2015-05-04 | 2016-11-10 | Dräger Safety AG & Co. KGaA | Head protection device, communication unit, connection unit and system comprising head protection device, communication unit and connection unit |
| CN206461605U (en) * | 2017-01-04 | 2017-09-01 | 泉州力同科技有限公司 | A kind of integrated talkback module and its system |
-
2018
- 2018-01-25 CN CN201810074257.XA patent/CN108417221B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1064562B1 (en) * | 1998-03-17 | 2009-01-14 | QUALCOMM Incorporated | System and method for determining the position of a wireless cdma transceiver |
| US8213953B1 (en) * | 2004-05-11 | 2012-07-03 | Sprint Spectrum L.P. | Method and system for vocoder bypass based on loads in coverage areas of a wireless wide area network |
| CN102763160A (en) * | 2010-02-18 | 2012-10-31 | 高通股份有限公司 | Microphone array subset selection for robust noise reduction |
| US20160249251A1 (en) * | 2015-02-19 | 2016-08-25 | Icom Incorporated | Radio communication system, radio apparatus, radio communication method, packet generation method, and method for reproducing data from packet |
| US20160329922A1 (en) * | 2015-05-04 | 2016-11-10 | Dräger Safety AG & Co. KGaA | Head protection device, communication unit, connection unit and system comprising head protection device, communication unit and connection unit |
| CN206461605U (en) * | 2017-01-04 | 2017-09-01 | 泉州力同科技有限公司 | A kind of integrated talkback module and its system |
Non-Patent Citations (5)
| Title |
|---|
| CARLOS URRUTIA-VALDÉS ET AL: "SIP in 3G Wireless Networks: Service Models,Architecture, and Network Design", 《BELL LABS TECHNICAL JOURNAL》 * |
| ZEBIN WU ET AL: "GPU Parallel Implementation of Spatially Adaptive Hyperspectral Image Classification", 《IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING》 * |
| 吴泽彬等: "基于迭代加权L1 正则化的高光谱混合像元分解", 《南京理工大学学报》 * |
| 韦建超: "基于AMBE声码器的数字对讲无线语音传输系统研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 黄慧猛: "基于数字对讲机的MELP声码器算法研究与实现", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108417221B (en) | 2021-09-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107748888B (en) | A kind of image text row detection method and device | |
| US20230177347A1 (en) | Method and apparatus for detecting defect pattern on wafer based on unsupervised learning | |
| CN105678612A (en) | Mobile terminal original certificate electronic intelligent filling system and method | |
| CN103034838B (en) | A kind of special vehicle instrument type identification based on characteristics of image and scaling method | |
| CN102324038B (en) | Plant species identification method based on digital image | |
| CN109241741B (en) | Malicious code classification method based on image texture fingerprints | |
| CN104834940A (en) | Medical image inspection disease classification method based on support vector machine (SVM) | |
| CN104036289A (en) | Hyperspectral image classification method based on spatial and spectral features and sparse representation | |
| CN110647951A (en) | Wireless radio frequency equipment identity recognition method and system based on machine learning algorithm | |
| CN106021578A (en) | Improved text classification algorithm based on integration of cluster and membership degree | |
| CN101256631A (en) | Method, apparatus, program and readable storage medium for character recognition | |
| CN104850859A (en) | Multi-scale analysis based image feature bag constructing method | |
| CN104751186A (en) | Iris image quality classification method based on BP (back propagation) network and wavelet transformation | |
| CN110458189A (en) | Compressed sensing and depth convolutional neural networks Power Quality Disturbance Classification Method | |
| CN111814825B (en) | Apple detection grading method and system based on genetic algorithm optimization support vector machine | |
| CN108376257B (en) | Incomplete code word identification method for gas meter | |
| CN115908142A (en) | Contact net tiny part damage testing method based on visual recognition | |
| CN107305640A (en) | A kind of method of unbalanced data classification | |
| CN104392244A (en) | Synthetic aperture radar image classifying method based on stacked automatic coding machines | |
| CN104732246B (en) | A kind of semi-supervised coorinated training hyperspectral image classification method | |
| CN111767909B (en) | Character recognition method and device and computer readable storage medium | |
| CN105868778A (en) | Method for detecting median filtering based on frequency domain fluctuation characteristics | |
| CN112200789A (en) | Image identification method and device, electronic equipment and storage medium | |
| CN108417221A (en) | The digital walkie-talkie acoustic code type detection method of fused filtering is recombinated based on signal two dimension | |
| CN116776245A (en) | Three-phase inverter equipment fault diagnosis method based on machine learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |